The machine learning model AutMedAI can predict autism in children under two with nearly 80% accuracy by analyzing 28 early-life parameters, such as the age of the first smile and eating difficulties. Developed by researchers at Karolinska Institutet, this model uses basic data to identify autism, which is crucial for timely intervention and improving developmental outcomes. The study, which used data from around 30,000 individuals, demonstrated that AutMedAI effectively detects autism-related difficulties and is a promising tool for early diagnosis. Further validation and the incorporation of genetic information are planned to enhance the model’s accuracy and clinical utility.
